Assessment of the Imaging Performance of the CITIUS High-Resolution Detector for Heavy Charged Particles and Neutrons
Pith reviewed 2026-05-25 06:25 UTC · model grok-4.3
The pith
CITIUS detector gain-selecting mode between high and medium gains improves spatial resolution for 4 MeV alpha particles from 9.1 μm to 1.2 μm and for cold neutrons from 26 μm to 1.9 μm at 70 μm pixel size.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Simulations based on a Geant4 model fitted to measured alpha-particle cluster shapes demonstrate that operating CITIUS in gain-selecting mode between high and medium gains produces substantial improvements in spatial resolution, reaching 1.2 μm for alpha particles and 1.9 μm for cold neutrons at a 70 μm pixel pitch, compared with 9.1 μm and 26 μm without gain selection; these gains arise because the long carrier drift distance enables substantial charge sharing while the dual-gain architecture preserves dynamic range.
What carries the argument
Gain-selecting architecture between high and medium gains together with long carrier drift distance that produces charge sharing across pixels.
If this is right
- The same two features—gain selection and long-drift charge sharing—extend the detector’s utility from X-rays to heavy charged particles and neutrons.
- At 70 μm pixel size the effective resolution for alphas improves by a factor of roughly 7.5.
- At the same pixel size the effective resolution for cold neutrons improves by a factor of roughly 14.
- The model allows quantitative prediction of performance at other pixel sizes and bias voltages once the four fitted parameters are fixed.
Where Pith is reading between the lines
- If the model holds, CITIUS could be used for position-sensitive neutron detection in scattering or radiography setups that currently require separate detectors.
- The approach suggests a general route for repurposing existing X-ray pixel detectors for charged-particle or neutron imaging by adding gain-selection logic.
- Testing the fitted model against a mixed beam of alphas and neutrons in one run would provide a low-cost internal consistency check.
Load-bearing premise
The four parameters fitted to alpha-particle data in the Geant4 model apply accurately to cold-neutron interactions without separate experimental confirmation.
What would settle it
A direct measurement of spatial resolution for cold neutrons incident on an actual CITIUS detector would show whether the simulated 1.9 μm figure at 70 μm pixels holds.
read the original abstract
We report on the assessment of the imaging performance of CITIUS -- a high-speed X-ray detector developed for the large-scale synchrotron radiation facility SPring-8-II -- for heavy charged particles and neutrons. To characterize the detector response, an irradiation experiment was performed using alpha particles from an $^{241}$Am source at four back-bias voltages of 400V, 300 V, 200 V, and 170 V, thereby controlling the amount of charge diffusion. A Geant4 model of the experiment was constructed, and four model parameters were determined by template fitting to the measured signal cluster shape distributions. The best-fit values are: an intrinsic energy spread of 5% for the source, a gold fraction of 0.4 for the Au-Pd coating, a lateral charge diffusion spread of 26.5 ${\mu}$m over a drift distance of 650 ${\mu}$m at 400V back-bias, and a per-pixel readout noise of 10000 $e^{-}$ in the medium-gain channel. Using the obtained sensor model, simulations were performed for 4 MeV alpha particles and cold neutrons to evaluate the expected spatial resolution. In both cases, simulated CITIUS, when operated in a gain-selecting mode between high and medium gains, yields a substantial improvement: at a pixel size of 70 ${\mu}$m for example, the resolution improves from 9.1 ${\mu}$m to 1.2 ${\mu}$m for alpha particles, and from 26 ${\mu}$m to 1.9 ${\mu}$m for cold neutrons. These results suggest that two key features of CITIUS -- its gain-selecting architecture and the substantial charge sharing enabled by the long carrier drift distance -- extend its imaging capabilities beyond X-rays to heavy charged particles and neutrons.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports an alpha-particle irradiation experiment on the CITIUS detector using a 241Am source at four back-bias voltages (400 V to 170 V). A Geant4 model is constructed and four parameters (5% intrinsic energy spread, 0.4 gold fraction in Au-Pd coating, 26.5 μm lateral diffusion over 650 μm at 400 V, and 10000 e- medium-gain noise) are obtained by template fitting to measured cluster shapes. These parameters are then used without modification to simulate 4 MeV alpha particles and cold neutrons, claiming that gain-selecting operation between high and medium gains improves resolution from 9.1 μm to 1.2 μm (alphas) and from 26 μm to 1.9 μm (neutrons) at 70 μm pixel size.
Significance. If the model transfer holds, the work would demonstrate that CITIUS's gain-selecting architecture and long-drift charge sharing can extend usefully to heavy charged particles and neutrons, providing a concrete performance prediction for applications in neutron imaging. The alpha experiment and template-fitting procedure supply direct experimental grounding for the sensor model; the simulation framework itself is reproducible in principle.
major comments (1)
- [Abstract and neutron-simulation results] Abstract and the neutron-simulation paragraph: the reported cold-neutron resolution improvement (26 μm → 1.9 μm at 70 μm pixels) is obtained by inserting the four alpha-fitted parameters unchanged into Geant4 runs for neutron-induced secondaries. Because the initial charge-cloud size, range, and dE/dx differ between 4 MeV direct alphas and neutron-capture products, and because the lateral-diffusion term was tuned specifically to the alpha drift geometry, the neutron prediction rests on an untested extrapolation; no neutron data or sensitivity test of the parameter transfer is shown.
minor comments (1)
- [Abstract] The abstract states the four best-fit values but does not give their uncertainties or the χ² of the template fit; adding these would allow readers to assess the robustness of the subsequent extrapolations.
Simulated Author's Rebuttal
We thank the referee for the detailed and constructive review. Below we address the major comment regarding the neutron simulation results.
read point-by-point responses
-
Referee: [Abstract and neutron-simulation results] Abstract and the neutron-simulation paragraph: the reported cold-neutron resolution improvement (26 μm → 1.9 μm at 70 μm pixels) is obtained by inserting the four alpha-fitted parameters unchanged into Geant4 runs for neutron-induced secondaries. Because the initial charge-cloud size, range, and dE/dx differ between 4 MeV direct alphas and neutron-capture products, and because the lateral-diffusion term was tuned specifically to the alpha drift geometry, the neutron prediction rests on an untested extrapolation; no neutron data or sensitivity test of the parameter transfer is shown.
Authors: The four parameters fitted from the alpha data describe detector-specific properties rather than particle-specific ones. The intrinsic energy spread applies to the 241Am source and is not used for neutrons. The gold fraction in the coating, the lateral diffusion (determined from drift distance and bias voltage), and the readout noise are all properties of the sensor and electronics that remain the same regardless of whether the charge is generated by alphas or by neutron-induced secondaries. Geant4 is used to compute the initial charge deposition for each case. We therefore consider the transfer justified on physical grounds. Nevertheless, we acknowledge that direct experimental validation with neutrons is absent. We will revise the text to state these assumptions explicitly and to include a sensitivity study in which the diffusion parameter is varied within the uncertainty obtained from the alpha fit. revision: yes
Circularity Check
No circularity: alpha-calibrated Geant4 model applied to neutron simulation is standard extrapolation, not reduction by construction
full rationale
The paper performs an alpha-particle irradiation experiment, fits four parameters (energy spread, gold fraction, diffusion spread, noise) via template matching to measured cluster shapes, then inserts those fixed values into Geant4 runs for both alphas and cold neutrons. The neutron resolution figures (e.g., 26 μm → 1.9 μm) are therefore simulation outputs under the transferred model, not quantities forced to equal the alpha fit by any equation or self-citation. No self-definitional loop, no fitted-input-called-prediction of a closely related observable, and no load-bearing self-citation or uniqueness theorem appears in the provided text. The transferability assumption is explicit and could be falsified by neutron data; the derivation chain remains self-contained against the alpha benchmark.
Axiom & Free-Parameter Ledger
free parameters (4)
- intrinsic energy spread =
5%
- gold fraction of Au-Pd coating =
0.4
- lateral charge diffusion spread =
26.5 μm
- per-pixel readout noise (medium gain) =
10000 e-
axioms (1)
- domain assumption Geant4 accurately models charge diffusion and collection in the sensor for both alphas and neutrons
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.